CHAPTER 9 REVIEW QUESTIONS

docx

School

University of Windsor *

*We aren’t endorsed by this school

Course

3080

Subject

Statistics

Date

Apr 3, 2024

Type

docx

Pages

Uploaded by GrandFire12610

CHAPTER 9 REVIEW QUESTIONS Q1) What is PRE measure? A1) PRE. measures tells us how much we can reduce our errors in predicting an outcomes if we know how two variables are linked. Q2) Suppose that the political party preferences of students are predicted from their fields of study. Would it be correct to use a lambda as a measure of association here? Why? What if anything would a lambda of .15 mean? A2) Lambda provides us with an indication of the strength of the relationship between independent and dependent variables. So, if you wanted to know the relationship between political party and field of study, then a lambda measurement would be accurate. A lambda of .15 would mean there is no relationship between the two variables. Q3) In statistics, what is the difference between positive or negative relationship? A3) In a positive relationship as scores on one variable rise so do scores on the other In a negative relationship as scores on one variable rise, scores on the other decline. Q4) What is the odds ratio? What are some of its issue? A4) Odds ratio (OR) is when we divide the odds for the first column by the odds for the second. The ratio of two odds. Issues for odds ratio is that it depends on dichotomies. Q5) Formula for odds ratio? A5) (a/c) / (b/d) = (a/c) x (d/b) = ad/bc Q6) Set up calculation to get Q. What is the relationship between Q and odds ratio? A6) To calculate Q, we take (ad-bc) / (ad+bc), wheres the OR we take ad/bc. Each is built up from the same elements, ad and bc, and in fact each is a straightforward function of the other Q= (OR-1) / (OR+1) Q7) What is the relationship between the odds ratio and the marginal totals? Between the odds ratio and "inner structure" of the table? A7) The stability of the OR and Q under changing marginals reflects their exclusive focus on the inner structure of a table, in the sense of the ratios amount the cell values within it. Q8) What are two ways to interpret Q? When does it equal gamma? A8) - Q can be interpreted as the excess of pairs favouring one type of association over those favouring the other as a proportion of all pairs favouring one or the other. - Q gives us the proportion by which we can reduce our errors in guessing whether pairs favour one form of association (or the other) if we know how the variables are linked

- For ordinal variables it is equivalent. For nominal variables we can interpret it without reference to ordering of the pairs. Q9) Give the formula for d and indicate what the various symbols in it refer to? A9) - d = C-D / C+D+Ty C = concordant D= discordant Ty = pairs of cases tied to the dependent variable Q10) What is the difference between gamma and d? A10) The change lies in the addition of Ty to the denominator. Q11) Why did Somers wish to suggest an alternative to gamma? A11) Somers believed that the effect of an independent variable should be assessed on all cases that it might have influenced. Any pair in which the cases differed on the independent variable was a possibility. However, pairs that differ on one variable but not the other are not used in gamma, so pairs tied on the dependent variable will be ignored even if they differ on the independent. Q12) Are gamma and d PRE measures? If so, what sorts of error do they look at? A12) Error 1, when we allow for the possibility of predicting cases have now tied to the dependent variable, but not the independent variable, becomes [C+D+Ty] /2. Error 2 consists of all pairs we guess wrong once we know whether concordant or discordant pairs are dominant. Q13) What is the difference between a symmetric and an asymmetric measure of association? Give an example of each. A13) - Asymmetric measure thats is one that changes values depending one which variable is seen as affecting which. (lambda) - Symmetric meaning that it gives the same value whichever way we view the variables. (gamma). Q14) Which of the statistics considered in this chapter is PRE? A14) lambda, gamma, Q (a special case of gamma), and Somers'd. Q15) Why do lambda, gamma, and d not have the same numeric values? A15) the necessary N depends greatly on the strength of association.

Q16) gamma = .618. What are two ways to intreat this value? A16) a value of .618 tells us that we can reduce our errors in predicting the direction of pairs by 61.8%. our value os .618 tells us that the excess of concordant over discordant pairs equal to 61.8 % Q17) d = .379. What are two ways to interpret this value? A17) - d gives us the excess of concordant pairs over discordant or vice versa as a proportion of all pairs distinguished on the independent variable. - d gives us the proportion by which we could reduce our errors in predicting the direction of pairs, if we had a precise measure of the DV, if the pairs created by this precise measurement were evenly divided into concordant and discordant, and if we knew how the two variables were linked.

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version